So i'm at a bit of a loss here.

I have a bunch of TLS 20w lamps. Single pin, instant start, no filaments. These things run according to Marcel on 57v and 370mA, so identical to conventional 20w TL.

A 40w one should run at about the same current, just at a higher voltage. But what do i see? The Stabilux light bulbs are specified at 500mA! This is confirmed by Marcel's pictures of his stash of Stabilux lamps.

Now i'm getting confused. How does this all work? Are the stabilux lamps explicitely meant to run on a lower than spec voltage and current? Then why didn't they stamp those lamps for that rating? It does not sound weird to underrun the incandescents to last longer, but then it makes no sense to call them 140/150v lamps, if they are intended to be run at 110v and last longer.

In the attachments you can find examples. Two 20w ones should be able to be used in series, so no more than 370mA should run. That adds up to a voltage of 114v. Add the 140v incandescent, and you end up at 254, with a mains voltage rating of 230v. That's underrunning it by 10%. That means 500mA - 10% = 450mA which is too much. Not even accounting for the nonlinear relation between voltage and current in an incandescent lamp.
Also, factor in that our mains voltage was allowed to be in the order of 244v back in the day with a norm of 220v, it means that almost the full 500mA could flow on moments that the mains voltage was particularly high.

The reason why i'm wondering about this, is because if i have 114v of the fluorescent tube, and i can find 115v light bulbs, i hope i can select one that makes the tubes operate at the appropriate 370mA while running the incandescents at maximum efficiency. 110-115v at 40w would get me *really* close to that 370mA in theory. 114v of the tubes + 110 or 120v of the light bulbs means i'm at close enough to 230v (my local mains) to run them.

Alternatively i have found a source of light bulbs of 150v or so in Czechia, but if i can get away with 110-115v light bulbs, that is likely a much more convenient source than the actual non-retail source of the 150v lamps. That source has a minimum order of 100pcs.

I'd start trying to run a single and a couple of 20W TLS tubes with a regular 230V 100W incandescent as a ballast. Next you can measure exact voltage drop across the incandescent ballast in both cases and also lamps current. Fluorescent lamp voltage drop is largely independent on lamp current within a reason. Next you can adjust proper lamp current by choosing different wattage of incandescents or even with some combinations like 100W + 15W. Do not forget to publish the results you get!

Again as it was discussed earlier, as lamp voltage is not sinewave, you can't just deduct ballast voltage as line RMS voltage minus lamp RMS voltage, much easier to get the result experimentally.

Not sure why ballast lamps were rated at 0.5A. Probably some old convention.

An interesting quirk is that while you can used reduced (comparing to full line) voltage incandescents as a ballast for self-starting tubes, this can not be done with a starter preheat circuit, as ballast lamp will burn off when the starter is closed.